In bandwidth limited wireless channels, such as QAM systems, higher order modulations can be used to increase data throughput. These higher order systems, however, are more sensitive to channel imperfections, especially nonlinear distortions encountered in practical channels, such as, for example, satellite channels. While there are several sources of nonlinear distortion, a major distortion source are high-power amplifiers (HPA) used at the transmitter. The characteristics of HPAs have nonlinear amplitude and phase characteristics that vary with time. Because of these nonlinear distortions, driving the HPA at or near the saturation power can become problematic. The effects of these nonlinearities may be decreased by using larger output back-offs such as operating in the quasi-linear region, but doing so reduces the output signal level.
A common technique used to minimize both the output back-offs and HPA nonlinearities is to pre-distort the transmitted signal prior to amplification. Pre-distortion performs an inverse of the HPA nonlinearity to make the post amplification signal as close to linear as possible. Two types of pre-distortion techniques have been used in the art: symbol-by-symbol pre-distortion and sample-by-sample pre-distortion. Both of these pre-distortion methods are open-looped inasmuch as neither method receives feedback from the receiving terminal.
In symbol-by-symbol pre-distortion, each modulated symbol is pre-distorted by a certain value. This pre-distortion value may be the average distortion experienced when a particular symbol is transmitted. The distortion for a symbol may also be based on its value and the value of the adjacent few symbols. The pre-distorted symbol may then be filtered, usually by a Square Root Raised Cosine (SRRC) filter for bandwidth shaping and then sent to the HPA for amplification. This pre-distortion method does not require any knowledge of the HPA characteristics.
In sample-by-sample pre-distortion, the digital values at the output of a shaping filter are typically pre-distorted by a certain value, called the pre-distortion coefficients, so as to make the overall channel quasi-linear. These pre-distortion coefficients are dependent on the HPA characteristics. Accordingly, sample-by-sample pre-distortion necessitates either prior knowledge of the HPA characteristics or estimating the HPA characteristics. Because this technique is tied to the specific characteristics of a particular HPA it can provide superior performance over the symbol-by-symbol technique.